| CPC H04N 19/186 (2014.11) [H04N 19/105 (2014.11); H04N 19/132 (2014.11); H04N 19/176 (2014.11); H04N 19/30 (2014.11); H04N 19/44 (2014.11); H04N 19/59 (2014.11)] | 15 Claims |

|
1. A method for decoding a video signal, comprising:
reconstructing a luma block corresponding to a chroma block, wherein the luma block is adjacent to a plurality of reconstructed neighboring luma samples, and wherein the chroma block is adjacent to a plurality of reconstructed neighboring chroma samples;
computing a plurality of down-sampled luma samples from the plurality of reconstructed neighboring luma samples;
identifying, from a sub-group of the plurality of computed down-sampled luma samples, two down-sampled maximum luma samples, wherein the two down-sampled maximum luma samples correspond to two first reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively, and the sub-group is composed of a predefined number of the computed down-sampled luma samples among the plurality of computed down-sampled luma samples;
identifying, from the sub-group of the plurality of computed down-sampled luma samples, two down-sampled minimum luma samples, wherein the two down-sampled minimum luma samples correspond to two second reconstructed chroma samples of the plurality of reconstructed neighboring chroma samples respectively;
averaging the two down-sampled maximum luma samples, the two down-sampled minimum luma samples, the two first reconstructed chroma samples, and the two second reconstructed chroma samples, respectively, to obtain an averaged down-sampled maximum luma sample, an averaged down-sampled minimum luma sample, an averaged first reconstructed chroma sample and an averaged second reconstructed chroma sample;
fitting a linear model through the averaged down-sampled maximum luma sample, the averaged down-sampled minimum luma sample, the averaged first reconstructed chroma sample, and the averaged second reconstructed chroma sample;
computing down-sampled luma samples from luma samples of the reconstructed luma block, wherein each down-sampled luma sample corresponds to a chroma sample of the chroma block; and
predicting chroma samples of the chroma block by applying the linear model to the corresponding computed down-sampled luma samples.
|